lanl / spiner

Performance portable routines for generic, tabulated, multi-dimensional data
https://lanl.github.io/spiner
BSD 3-Clause "New" or "Revised" License
7 stars 3 forks source link

First commit to cleanup attributes and add forceinline to portable ar… #13

Closed dholladay00 closed 3 years ago

dholladay00 commented 3 years ago

…ray call operator accessors.

Addresses issues #12 plus additional cleanup

PR Summary

Eliminates the use of __attribute__s except in one macro definition.

PR Checklist

dholladay00 commented 3 years ago

@Yurlungur this should probably go through a performance analysis check to make sure there are no regressions in performance. You might be the most equipped to handle that.

Yurlungur commented 3 years ago

Hmm. Yes, I'll try to take a look. Some simple timing on, say, a convergence test, is probably sufficient for now.

dholladay00 commented 3 years ago

Tested with Kokkos with an OpenMP and Cuda backend. Saw no difference in performance from Main. I think we're good. Merge at will.

I'm not authorized to merge @Yurlungur, can you click the button?

Yurlungur commented 3 years ago

Done. And you should probably have merge priveleges. Let me add that.